741548 - .svg files served with wrong Content-Type

Status: RESOLVED FIXED

(Infrastructure & Operations Graveyard :: WebOps: Other, task)

Description

[Gregory Szorc [:gps]]

http://docs.services.mozilla.com is serving .svg files with Content-Type: text/xml. No web browser will render <img> SVG files served with this media type.

Please change the Apache server to serve .svg files with:

  Content-Type: image/svg+xml

Here is a test URL: http://docs.services.mozilla.com/_images/graphviz-e85bf880bff15f660f93361d9ea5c500e0e6ddf2.svg

Comment 1

[Gregory Szorc [:gps]]

Test rendering within an HTML page from http://docs.services.mozilla.com/sync/storageformat6.html

Comment 2

[:Atoll]

:gps, once 741564 is resolved in the next 7 days, this bug may go away. follow up with me once 741564 is resolved.

Comment 3

[Gregory Szorc [:gps]]

(In reply to Richard Soderberg [:atoll] from comment #2)
> :gps, once 741564 is resolved in the next 7 days, this bug may go away.
> follow up with me once 741564 is resolved.

It has been 10 days. Ping?

Comment 4

[:Atoll]

No word from Infra on 741564, so going with plan B, alter the existing httpd configs.

I looked into doing that. If I misunderstand and my change is erroneous, I risk crashing the entire Infra static cluster.

I will try to find someone in Webops tomorrow to confirm my changes before I commit them, or at worst, reassign this to Webops.

Ping after Tuesday if no word by then.

Comment 5

[:Atoll]

(In reply to Gregory Szorc [:gps] from comment #0)
> http://docs.services.mozilla.com is serving .svg files with Content-Type:
> text/xml. No web browser will render <img> SVG files served with this media
> type.
> 
> Please change the Apache server to serve .svg files with:
> 
>   Content-Type: image/svg+xml

Hi, IT. We were hoping the move out of SJC1 would fix this. It did not, as apparently PHX1 also ships a broken SVG config. Could you please repair the SVG content-type for the site "docs.services.mozilla.com"?

I have insufficient information about this Webops-run cluster to safely make this change without putting other sites at risk, so I'm handing this bug to the triage queue for triage.

Comment 6

[Jake Maul [:jakem]]

This is fixed!